IndyRadar | Live-ops dashboard for Aura & CRDB

Monitors Neo4j Aura and CockroachDB clusters to flag orphaned/duplicate instances and cost/health anomalies in real time.

Lead Product Designer & Frontend Engineer | 2024 - 2025

Snapshot

What: Internal live-ops web app unifying Neo4j Aura + CockroachDB signals for management and ops.

Who: Ops leadership, SRE/DBA, engineering managers.

My Role: Discovery → UX/UI → React/Next.js build (dashboards, auth, state), plus CI/CD and release hygiene.

Stack: Next.js 14, TypeScript, NextAuth (Google OAuth), SWR, OpenAPI/REST, Neo4j Aura + CRDB, GCP Cloud Run/Build.

Outcomes: Lighthouse 95+ in production • Issues surfaced before customers • Faster triage via real-time error/cost views.

Outcomes

  • Shipped at Lighthouse 95+ in production.
  • Issues surfaced before customers via real-time error and cost views.
  • Faster triage for ops and leadership; single source of truth across Neo4j Aura and CRDB.

Problem → Insight

  • Problem: Hidden orphan/duplicate instances and fragmented signals created blind spots and delayed response.
  • Insight: A single, low-latency dashboard with opinionated views and alerts would make risk and cost visible and actionable for ops and managers.

What I did (design + build)

  • Drove discovery with ops and management; defined top questions and alert thresholds.
  • Designed dashboards for instance health, costs, and anomalies; prioritized signal-to-noise.
  • Built Next.js 14/TypeScript UI: auth with NextAuth (Google OAuth), data fetching with SWR, OpenAPI/REST integration, guarded routes.
  • Set up CI/CD on GCP Cloud Build/Run; smoke tests and performance budgets; release hygiene.

How it works (compact flow)

  • Google OAuth via NextAuth; roles gate access.
  • API layer aggregates Aura and CRDB metadata and metrics.
  • SWR keeps views live; anomaly rules highlight risk and spend.
  • Users drill from fleet to instance; actions link to source systems.
  • CI checks perf and critical flows before deploy.

Selected screens

IndyRadar fleet overview showing Aura and CRDB tiles with risk and spend.

Fleet overview: risk and spend across Aura + CRDB in one view; drill into any tenant or cluster.

Grid listing orphan and duplicate instances across Aura and CRDB with flags.

Unified orphan/duplicate records: cross-store matches flag hidden risks and wasted cloud spend before customers notice.

Matched instances table with status, customer grouping, and actions.

Matched instances: status badges and grouping reveal duplicates instantly; triage starts from the row.

Instance detail panel showing health, error log, and cost trend.

Instance detail: health, errors, recent changes and cost trend in one place—designed for fast, confident decisions.

IndyRadar sign-in with Google OAuth and role badge.

Google OAuth + roles: access is gated by org role; sensitive routes guarded in UI and API.

Trade-offs & constraints

  • Started with REST/OpenAPI for speed; postponed graph queries beyond simple lookups.
  • Alerting remains informational in v1; automated remediation intentionally deferred.

What’s next

  • Alert subscriptions and weekly digest for leadership.
  • Cost attribution per team/project; spend guardrails.
  • Deeper graph views for lineage and dependency impact.